Posted by: oikosasa | April 2, 2013

Large-scale testing of Bergman’s rule

The colder, the bigger, suggested Bergman in 1848. In 2013 we publish a paper testing Bergman’s rule on a large data set. Showing…well find out in “Bergmann′s rule in mammals: a cross-species interspecific pattern” by Marcus Clauss and his co-workers. Below is their background story to the study:

I first learnt about ‘Bergmann’s rule’ (that among closely related species, those living at higher latitudes/at colder temperatures are larger) in school. It was one of the biological facts I had always considered a background fact that is unquestioned.  

When preparing a manuscript on the reproductive seasonality of ruminants (Zerbe et al., 2012) we collated various biological data on ruminants, including body mass and mid-latitude of their geographical range, and before testing relationships of these data with our proxy for seasonality, we tested them amongst themselves for potential correlations. We did this without accounting for the phylogenetic structure of the data (ordinary least squares) and with such an accounting (phylogenetic generalized least squares, pgls).


We were not surprised when we found a relationship between latitude and body mass in our phylogenetic analysis – because this simply reflected Bergmann’s rule. The fact that this relationship was not significant in conventional statistics, but significant in pgls, just supported the notion that the rule holds among closely related species, not just any species you lump together. For us, this was a minor side-dish result in our set of analyses, and one we were not excited about, because it simply confirmed what we knew from school. In a quite similar way, our ruminant dataset supported, for example, Rensch’s rule (which was cool for our co-author that had the same name). But when we prepared the manuscript, and searched for other papers to cite in connection with our side result, we realized that Bergmann’s rule had, as far as we could find, not been analysed in this fashion among mammal species, not in ruminants, and surely not in a larger mammal dataset. Literature on Bergmann’s rule in mammals most often dealt with the intra-specific side of the phenomenon or with mammal assemblages, but not on the taxonomic/interspecific level. So we expanded the dataset beyond the ruminant species for which we had seasonality data, to comprise all mammals (based on availability of data in the PanTheria database). Again, we found the same effect: the relationship between latitude and body mass was significant if the phylogenetic structure of the dataset was taken into account. In a sense, we felt like having found a simple proof for a school lesson that had not been provided so far. This does not mean we claim to be first, best, closest, whatever, to proving Bergmann’s rule – we just found a simple, maybe elegant way to demonstrate it. Actually, once you start looking at individual taxonomic subgroups or certain geographic ranges, the picture becomes less simple – but that’s in the paper.  

For us, several lessons came with working on this topic. One is that the statistical procedure (pgls, sometimes called ‘phylogenetically controlled statistics’) is not only a quarrelsome one that one has to apply nowadays to get a paper published, but actually facilitates, in special circumstances, the detection of a pattern that would not be evident from simply plotting the data, or from ‘conventional’ statistics. So we tried to understand how patterns would look that yield different results in conventional and phylogeny-informed statistics, drawing on textbooks on the matter, and produced some schematic graphs to get a mental grasp on this (provided as a supplement to our paper). In my experience, there are quite some examples where a relationship that is significant in conventional statistics is no longer so in phylogenetic statistics (which then needs to be interpreted), but there are very few examples where a relationship is not significant in conventional stats but clearly is in phylo-stats.  

The other lesson came from going through the literature – at some stage, we decided to try to locate the original source itself, Bergmann’s own account, which was sometimes described as ‘hard to get at’ – and were surprised that you could simply download the whole text from the net (it is now part of the google books resource pool). Reading this text was quite some fun, due to the old style it was written in. From other literature, we had gained the impression that Bergmann formulated his rule for mammals, and were therefore surprised to see that he actually developed it, and supported it, using birds. And all the examples Bergmann used himself were between, not within, bird species, so no need to debate whether he meant it on an inter- or intra-specific level. Because the text is written in German, we decided to provide a relatively detailed translation of larger parts of it, so that others could get a picture of how he built his argument. I personally especially cherished his concluding comment, where he cautions the reader that in trying to support his hypothesis, he might have looked at the actual evidence in a biased way, and that therefore independent tests would be welcome.    

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s


%d bloggers like this: